ABSTRACT
This chapter presents the method of estimation of the local signal bandwidth based on counting the level crossings. It provides a comprehensive introduction to the problem of estimation of the intensity function from a given set of level-crossing samples. The chapter describes the concept of the local bandwidth and the corresponding signal recovery problem. It explains the methodology for level-crossing sampling for signals that are modeled as Gaussian stochastic processes. The chapter discusses the signal reconstruction method for non-uniformly distributed samples that occur in the level-crossing sampling strategy and any other schemes with irregular sampling. It suggests that the varying bandwidth signal represented using time-warping method can be transformed into the constant bandwidth signal. The well-designed event-triggered sampling system provides the number of samples proportional to the bandwidth of a signal. Level-crossing sampling is an eventbased sampling scheme providing the samples whose density is dependent on the local bandwidth of the sampled signal.
